rdtr opened a new pull request, #13564: URL: https://github.com/apache/gluten/pull/13564
## What changes are proposed in this pull request? `GlutenParquetTypeWideningSuite` excludes 13 `parquet widening conversion` tests in the spark41 and spark42 `VeloxTestSettings`. They write an integer column (`ByteType`, `ShortType`, `IntegerType` or `LongType`) and read it back as a wider `DecimalType` or as `DoubleType`, for example `IntegerType -> DecimalType(20,0)` and `ShortType -> DoubleType`. The same 13 excludes were removed from spark40 by the Velox update in #11860, but spark41 and spark42 kept them. They pass now on both versions, so this removes the 26 excludes. The `TimestampNTZType` widening excludes and the decimal precision and scale excludes stay. The reads run in Velox. For a Parquet file with an `INT` column read with `a DECIMAL(20,0)` or `a DOUBLE`, the executed plan on Spark 4.2 is `VeloxColumnarToRow` over `FileFileSourceScanExecTransformer parquet [a#5076] ... ReadSchema: struct<a:decimal(20,0)>`. ## How was this patch tested? Ran `GlutenParquetTypeWideningSuite` with `SPARK_TESTING=true`, `SPARK_ANSI_SQL_MODE=false`, `-Pdelta` and `-Dspark.test.home` set, as in the CI jobs: - Spark 4.1: 78 succeeded, 0 failed. - Spark 4.2: 78 succeeded, 0 failed. ## Was this patch authored or co-authored using generative AI tooling? Generated-by: Claude Code (Claude Opus 5.5) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
